diff --git a/docs/implementation.md b/docs/implementation.md
index ed6b82f..5b5f561 100644
--- a/docs/implementation.md
+++ b/docs/implementation.md
@@ -141,3 +141,14 @@ current packages, app/website commands, verification previews, analytics limits,
 skill location and contribution workflow. Local links and diff checks pass;
 commands and prerequisites were checked against repository configuration.
 Documentation only; no runtime tests were run.
+
+## UX simplicity preference
+
+Task [#26](https://git.sendico.io/tech/sxchnge/issues/26), delivered in
+[PR #27](https://git.sendico.io/tech/sxchnge/pulls/27) on `skills-26`, makes
+simplicity an explicit preference in the UX Design skill and its review checklist.
+Prefer fewer concepts, choices, and controls when they serve the same need; retain
+essential information, accessibility, and recovery. Documentation only.
+Validation: skill-creator `quick_validate.py` passed in an isolated environment
+with PyYAML (the system Python lacked it); `git diff --check` passed. No runtime
+tests were needed. The PR is not merged.
diff --git a/skills/ux-design/SKILL.md b/skills/ux-design/SKILL.md
index 9a72d17..1645605 100644
--- a/skills/ux-design/SKILL.md
+++ b/skills/ux-design/SKILL.md
@@ -10,6 +10,10 @@ before reporting it complete. Rendering correctly is only one part of completion
 
 ## Establish the flow
 
+- Prefer the simplest interface that lets users complete their task confidently.
+  When options meet the same need, choose fewer concepts, choices, and controls.
+  Add complexity only for a concrete user need; preserve essential information,
+  accessibility, and recovery rather than hiding them to make a screen look minimal.
 - Identify the user's primary task, successful outcome, and shortest appropriate
   path to that outcome before choosing layouts or controls.
 - Inspect comparable product flows, components, navigation, terminology, and
@@ -95,6 +99,8 @@ performed and any gaps; do not mark applicable unmet criteria as complete.
 - [ ] Are interactive elements recognizable, and can the user tell what is
   happening and what will happen next?
 - [ ] Are steps appropriate, with unnecessary interaction and cognitive load removed?
+- [ ] Is this the simplest usable solution, with each added concept, choice, or
+  control serving a concrete user need?
 - [ ] Does the flow follow existing product patterns and consistent terminology,
   with a concrete reason for any departure?
 - [ ] Can users recover from errors, with input and context preserved where appropriate?